United States Department of Agriculture

Transcription

1 United States Department of Agriculture Historical and Current Fire Management Practices in Two Wilderness Areas in the Southwestern United States: The Saguaro Wilderness Area and the Gila-Aldo Leopold Wilderness Complex Molly E. Hunter, Jose M. Iniguez, and Calvin A. Farris Forest Rocky Mountain General Technical Report Service Research Station RMRS-GTR-325 August 2014

2 Hunter, Molly E.; Iniguez, Jose M.; Farris, Calvin A Historical and current fire management practices in two wilderness areas in the southwestern United States: The Saguaro Wilderness Area and the Gila-Aldo Leopold Wilderness Complex. Gen. Tech. Rep. RMRS- GTR-325. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 38 p. Abstract Fire suppression has been the dominant fire management strategy in the West over the last century. However, managers of the Gila and Aldo Leopold Wilderness Complex in New Mexico and the Saguaro Wilderness Area in Arizona have allowed fire to play a more natural role for decades. This report summarizes the effects of these fire management practices on key resources, and documents common challenges in implementing these practices and lessons for how to address them. By updating historical fire atlases, we show how fire patterns have changed with adoption of new policy and practices. Keywords: prescribed fire, managed wildfire, Southwest Authors Molly E. Hunter, School of Forestry, Northern Arizona University, Flagstaff, Arizona. Jose M. Iniguez, USDA Forest Service, Rocky Mountain Research Station, Flagstaff, Arizona. Calvin A. Farris, National Park Service, Klamath Falls, Oregon. You may order additional copies of this publication by sending your mailing information in label form through one of the following media. Please specify the publication title and series number. Telephone: (970) FAX: (970) rmrspubrequest@fs.fed.us Website: Mailing address: Publications Distribution Rocky Mountain Research Station 240 West Prospect Road Fort Collins, CO 80526

3 Contents INTRODUCTION... 1 Background, Purpose and Scope...1 Terminology...2 METHODS... 2 STUDY AREAS... 3 Geography...3 Vegetation Communities...5 Cultural History...8 HISTORY OF FIRE REGIMES AND MANAGEMENT... 9 Pre-Euro-American Settlement Era (Prior to 1900)...9 Fire Exclusion Era ( ) Managed Fire Era ( )...12 Megafire Era (2011 present)...17 EFFECTS OF MANGED WILDFIRE AND PRESCRIBED FIRE Fuels, Forest Structure, and Potential Fire Behavior...20 Threatened, Endangered, and Sensitive Species...23 Aquatic Habitat and Species...23 Mexican Spotted Owl...23 Wildlife Habitat and Populations...24 CURRENT CHALLENGES TO IMPLEMENTING MANAGED WILDFIRE Threatened and Endangered Species...25 Public Support for Fire Programs...25 Risk of Escape...26 Invasive Species...26 Reintroducing Fire in Mixed to High Severity Fire Regimes...27 LESSONS LEARNED Managed Wildfire and Prescribed Fire Can Restore Southwestern Forests...28 Managed Wildfire, Large-Scale Prescribed Fire, and Suppression Fires Create Landscape Fire Mosaics That Provide Opportunities for Fire Managers...29 Collaboration Across Disciplines and Agencies Is Imperative...29 Preparation and Public Outreach Is Critical...30 CONCLUSION ACKNOWLEDGMENTS REFERENCES GLOSSARY APPENDIX A: MANAGERS INTERVIEWED APPENDIX B: RESEARCH NEEDS... 37

4

5 INTRODUCTION Background, Purpose and Scope Suppressing wildfires has been the dominant fire management strategy over most of the western United States for the last century. In forested systems adapted to frequent fire, this practice, in part, has resulted in contemporary forests with declining ecological conditions, large build-up of fuels, and increased potential for large, high-intensity and stand replacing wildfire (Covington and others 1994). However, in some areas managers have been able to adopt more flexible policies over the last few decades to allow fire to play a more natural role. Two of the most notable examples occur in the southwestern United States: the Gila and Aldo Leopold Wilderness Complex (GALWC) in the Gila National Forest (GNF) in New Mexico, and the Saguaro Wilderness Area (SWA) in Saguaro National Park (SNP) in Arizona (figure 1). Over the last 30 to 40 years, policies that facilitated the management of lightning and prescribed fires in these areas have resulted in the large-scale reintroduction of frequent fire as an ecological process. As a result, current forest conditions are strikingly different compared to long-unburned forests that can be found throughout the rest of Southwest. Forest managers across the Southwest have long recognized the need to restore historical forest structure and reduce fuels following a century or more of fire exclusion. Mechanical treatments are often used to accomplish this goal in non-wilderness areas, but managers are increasingly relying on fire as a restoration tool for three main reasons: (1) the lack of a timber industry and local wood products manufacturing plants that can make mechanical treatments economically viable, (2) recent fire policy guidelines that now allow more flexibility in the use of fire, and (3) the need to work at landscape scales to effectively reduce the threat of large, high-intensity fires. Fire is often more efficient to implement than mechanical methods across large scales; the cost per acre of fire is typically lower and it is less restricted by access, topography, or legal challenges (Hunter and others 2007). Similar circumstances initially propelled the use of large-scale fires in wilderness areas within the GALWC and SWA. After several decades of fire management these two areas offer a wealth of knowledge that could aid managers who are now beginning to rely more on fire as a restoration tool in other forests throughout the western United States. The purpose of this report is to document and synthesize the current and historical fire management practices in the GALWC and SWA, the effects of reintroduced fires on critical natural resources, and provide examples of how managers have addressed common challenges in implementing these fire programs. We culminate our findings with a discussion of key lessons learned from these unique programs. These lessons are broad concepts that are supported by the literature and conversations with managers and are applicable across both study areas. They represent what we have learned from these fire management programs that could be applicable to other areas where the practice of using fire as a restoration tool at landscape scales is in its infancy. Figure 1 Location of the Saguaro Wilderness Area in Saguaro National Park (Arizona) and the Gila Aldo Leopold Wilderness Complex in the Gila National Forest (New Mexico). USDA Forest Service RMRS-GTR

6 Terminology Terminology regarding the management of naturally ignited fire for restoration has changed numerous times since the practice was first adopted as official policy. When the policy was first implemented by a handful of National Park units in the late 1960s it was called a variety of things: let burn, natural prescribed fire, or natural fire management (van Wagtendonk, 2007). Over time, as policies were revised, the term prescribed natural fire took hold. A review and update of national fire policy in 1995 further reaffirmed the practice and introduced the term wildland fire use for resource benefit, referred to as wildland fire use or often fire use (Philpot and others 1995). The latest change in terminology came with the release of the 2009 Federal Interagency Implementation Guide (Fire Executive Council 2009). In the past, official policy distinguished three types of fire: wildland fire use, prescribed fire (intentionally ignited by managers), and wildfire (unwanted and suppressed fire). The most current guidelines recognized just two types of fire: wildfire and prescribed fire. This is an important distinction because under the former policy a fire resulting from a natural ignition had to be declared either a wildfire or wildland fire use and either suppressed or managed for resource objectives (but not both). With the new policy interpretation, a single wildfire can be managed for multiple objectives, allowing mangers to aggressively suppress areas of a fire where resources are threatened and apply minimal suppression strategies in areas where a fire is benefiting and/or posing minimal threat to key resources. A common terminology describing this practice has yet to be widely adopted and can include multiple labels, including, minimal suppression, contain, confine, and multiple objective fire. For simplicity, throughout this report we use the term managed wildfire to describe all naturally ignited fires in GWALC and SWA that have been managed with a strategy other than full suppression. This would encompass any fires that in the past might have been labeled prescribed natural fire, wildland fire use, multiple objective fire, or other related terms. Regardless of the specific terminologies used over time, these natural ignitions were managed with relatively similar overall strategies and objectives, and can be lumped into a single category. The term prescribed fire is used to describe any fire intentionally set by managers to meet specific resource objectives and suppression fire to describe any unwanted fire that was actively suppressed. It should be noted that there is obviously overlap between managed, prescribed, and suppression fires. For example, a full suppression fire may involve indirect tactics such as utilizing natural barriers for safety. Conversely, managed wildfires often involve some level of direct suppression strategies such as line construction and firing operations to limit spread into undesirable areas or to manipulate fire intensity and effects (particularly in recent years due to more flexible policies). Moreover, some fires in the past started under one classification but were converted to another. For example, the 1994 lightning-ignited Rincon Fire in SWA was initially a managed wildfire (prescribed natural fire at that time) but was subsequently converted to full suppression fire when it burned outside of a specified area and acceptable fire behavior/weather parameters. Despite these ambiguities, this classification serves as a useful means of clarifying confusing terminology and examining how the use of fire has changed over time in the study areas. METHODS We used a combination of approaches to document and synthesize the historical role of fire and the effects of contemporary fire programs on critical resources in the GWALC and SWA. First, we conducted an extensive literature review of both peer-reviewed studies and grey literature regarding fire ecology and fire management practices in both study areas. Secondly, to identify challenges and keys to success in implementing the fire programs, we interviewed several managers (both current and retired) from a variety of disciplines in each study area. This included phone interviews, discussions, and informal conversations on field trips and in other settings. Those interviewed represent over 200 years of cumulative fire management experience in these study areas; they provided invaluable insight into the challenges and processes of implementing innovative fire management strategies (see list of names in Appendix A). Results from that effort are documented throughout this report. Third, to document the landscape scale patterns of fire occurrence over time resulting from changing management strategies, we updated previously published fire history atlases (Farris and others 2010; Rollins and others 2001) using data obtained from the National Park Service, The U.S. Forest Service, and the Monitoring Trends in Burn Severity database ( last accessed 6/7/2013). Fire frequency maps encompassed the period from 1909 to 2013 at GALWC, and 1937 to 2013 at SWA, and included adjacent areas. Prescribed 2 USDA Forest Service RMRS-GTR

7 fires were not listed in the original fire history atlas for the GNF. However, using the Monitoring Trends in Burn Severity database ( we were able to update it to include prescribed fires over 1,000 acres in size that occurred after We also calculated cumulative area burned and proportion of area burned by each fire type (managed, prescribed, suppression) over time. In the SWA, this analysis was restricted to the greater Mica Mountain Study Area (MMSA). This area supports the primary coniferous forest belt in the wilderness, which has been the main focus of most managed and prescribed fires. We conclude the report with some key lessons learned from the fire management programs in the GAWLC and SWA. These were the culmination of the work that went into producing this report (literature review, interviews with managers, and fire history analysis). The list of lessons learned is certainly not comprehensive, but it represents a set of recurring, core ideas that were conveyed independently by various managers across both of the study areas, and had strong support in the literature and in our own analyses. We chose lessons with multiple lines of corroboration, assuming that they would be more likely to be applicable across broad geographic areas, including non-wilderness, and potentially most useful where using fire as a large-scale restoration tool is at an earlier stage of development. We also present these lessons as a starting point in hopes that managers in other areas will add to them based on their own experience. STUDY AREAS Geography The GALWC in central New Mexico encompasses 558,014 acres and is composed of two main mountain ranges (figure 2). The Mogollon Mountains dominate the western half of the study area. They rise to an elevation of 10,742 feet and fall abruptly toward the west and south and more gently toward east and north. The Black Range is a smaller range that runs south to north along the eastern part of the GALWC. These two main mountain ranges are the headwaters for the Gila River. Between them the land is dominated by steep canyons and large expansive mesas. The entire landscape was formed by volcanic activity in the Cretaceous Era (USGS 1965). The SWA in southern Arizona is 70,905 acres and encompasses most of the Rincon Mountains, which are part of the Basin and Range Province that dominates the landscape across southern Arizona and the northern state of Sonora in Mexico. These mountain ranges were uplifted during the mid-tertiary Era (Baisan 1990). The main mass of the Rincon Mountains is composed of granite gneiss although limestone outcrops can also be found throughout. The Rincon Mountains rise abruptly from the Sonoran Desert to an elevation of 8,432 feet above sea level. The two main physical features of the Rincon Mountains are two peaks, Mica Mountain and Rincon Peak to the south (figure 3). Mica Mountain Figure 2 Detailed map of the Gila Aldo Leopold Wilderness Complex and surrounding lands in the Gila National Forest. The main physical features of the Gila Wilderness are the Mogollon Mountains in the northwest along with the three forks of the upper Gila River. The main physical feature of the Aldo Leopold Wilderness is the north-south running Black Range. USDA Forest Service RMRS-GTR

8 Figure 3 Detailed map of the Saguaro Wilderness Area in Saguaro National Park and the adjacent Rincon Mountain Wilderness Area in the Coronado National Forest. The outlined Mica Mountain Study Area represents a 6,870 acre area on Mica Mountain where the prescribed fire and managed wildfire program has been focused. It is also where most fire history studies have been conducted. slopes gently along the top but drops off abruptly on three sides towards the desert. Rincon Peak is steeper and dissected. The two high points are connected by Happy Valley saddle, which supports a stand of ponderosa pine at 6,234 feet elevation (figure 4). Both study areas have a seasonally bimodal precipitation pattern. In the winter (November to March), frontal storms bring snow to higher elevations and rain to the lower elevations. In the mid- to late-summer (July to August), moisture from Mexico creates a monsoonal weather pattern with thunderstorms. The arid foresummer period (April to June) preceding the monsoonal moisture is typically very hot and dry and is when most large fires burn. Annual precipitation varies greatly by elevation. In the GAWLC, averages range between 12 to 28 inches from roughly 4,500 to 11,000 feet in elevation. In the SWA, average annual precipitation ranges from about 13 inches at the base of Mica Mountain (~2,500 feet) to 27 inches at the summit (8,666 feet). Both study areas represent very remote and inaccessible areas, making them ideal places to experiment with managed wildfire. While there are some towns on the outskirts of the GALWC and GNF (Silver City, New Mexico, and Reserve, New Mexico) and private inholdings throughout, the human population density in the region is relatively low. Although SWA and SNP is located next to a major metropolitan area (Tucson, Arizona), its rugged, roadless terrain, difficult access, and surrounding unburnable fuels (desert) resulted in functional isolation for fire management purposes (figure 5). While there are some historical structures and significant cultural sites in both study areas, there are Figure 4 View from the top of Rincon Peak looking north across Happy Valley Saddle to Mica Mountain in the background (photograph by Michael Crimmins, University of Arizona). 4 USDA Forest Service RMRS-GTR

9 Figure 5 View of ponderosa pine and mixed conifer forests in the Mica Mountain Study Area (Helens Dome in foreground), looking west across desert vegetation and the city of Tucson Arizona in the background (photograph by Calvin Farris, National Park Service). very few modern human developments directly adjacent to the wilderness areas. Vegetation Communities Both study areas are located at the junction of two major biomes. From the north there is a Rocky Mountain Cordillera influence and from the south there is a Madrean influence. The sharp vertical relief and unique geographic position of these mountains combine to create a diverse mosaic of vegetation assemblages stratified along elevation/moisture gradients (Niering and Lowe 1984). Plant communities at higher elevations (above 6,000 feet) are closely related to the Cordilleran/Rocky Mountain floristic province. Plant communities in lower elevations (below 6,000 feet) are primarily influenced by the Madrean floristic province (McLaughlin 1995). For the sake of simplicity, however, we have identified three broad vegetation communities that this report will focus on: pinyon-juniper woodlands, ponderosa pine forests, and mixed-conifer forests (figure 6). Pinyon-juniper woodlands typically occur in areas below 6,500 feet elevation although they can also be found at higher elevations on drier south-facing aspects (figure 7). The dominant tree species within the pinyon-juniper plant community include alligator juniper (Juniperus deppeana Steud.), Rocky Mountain juniper (Juniperus scopulorum Sarg.), one-seed juniper (Juniperus monosperma (Engelm.)Sarg.), pinyon pine (Pinus edulis Engelm.) and border pinyon (P. discolor D.K. Bailey and Hawksw). Three general pinyon-juniper vegetation types have been identified across the western United States: persistent pinyon-juniper woodlands, pinyon-juniper savannas, and wooded shrublands (Romme and others 2009). The persistent woodland and wooded shrubland types are generally characterized as having higher tree and shrub density and lower grass cover in the understory compared to the savanna type. Both the persistent woodland and savanna types occur in the GNF, although the savanna type was historically much less prevalent (Miller 1999). Various pinyon and juniper species can be found at lower elevations in the pine/oak woodlands in SWA, which also includes a number of Madrean oak species. Ponderosa pine forests are generally found at midelevations between 6,500 and 8,500 feet elevation (figure 8). The dominant tree species in these forests are ponderosa pine (Pinus ponderosa C. Lawson) and sometimes Gambel oak (Quercus gambelii Nutt.). In SWA, pine forests also includes Madrean species such as Apache pine (Pinus engelmanii Carr.), Chihuahua pine (P. leiophylla Schiede & Deppe), Arizona pine (Pinus arizonica), silverleaf oak (Quercus hypoleucoides A. Camus), Arizona white oak (Q. arizonica Sarg.), and Emery oak (Q. emoryi Torr.) (figure 9). Mixed conifer forests are generally found between 8,500 and 10,000 feet elevation (figure 10). Common tree species within mixed conifer forests include Douglas-fir (Pseudotsuga menziesii [Mirb.] Fanco), white fir (Abies concolor [Gord. & Glend.] Hildebr.), southwestern white pine (Pinus strobiformis Engelm.) quaking aspen (populus tremuloides Michx.) and ponderosa pine. Forest structure and composition generally changes dramatically along an elevation-temperaturemoisture gradient between warmer, drier ponderosa pine and colder, wetter spruce-fir forest types. Mixed-conifer can be further separated into warm/dry and cool/moist subtypes that occupy each end, respectively, of the elevation-temperature-moisture gradient (Evans and others 2011) (figure 11). The cool/moist type occurs at the highest elevations but can also be found on north-facing aspects at lower elevations (figure 11a). The main distinguishing feature of this type is that although ponderosa pines maybe present they are not usually dominant. The warm/dry type occurs at the lower elevations or on south-facing slopes at higher elevations. This type can USDA Forest Service RMRS-GTR

10 Figure 6 Maps of vegetation communities in (a) the Gila and Aldo Leopold Wilderness Complex and surrounding lands, and (b) Saguaro National Park. Vegetation categories are derived from Landfire Data ( last accessed 1/21/14). 6 USDA Forest Service RMRS-GTR

11 Figure 7 Typical unburned pinyon-juniper vegetation in the Gila National Forest. This vegetation community tends to occur at lower elevation within these mountain ranges (photograph by Molly Hunter, Northern Arizona University). Figure 9 Ponderosa pine forests in Saguaro National Park often include a number of Madrean oak species in the understory. Rincon Peak is in the background (photograph by Calvin Farris, National Park Service). Figure 8 Southwest ponderosa pine forest in the Gila National Forest. Although ponderosa pine can be found with other species at its lower and higher elevation limits, it often forms large stands or pure ponderosa pine. Historically, such stands were open and dominated by large trees due to frequent fires; however, due to logging and fire suppression, stands such as this one in the Gila wilderness are relatively rare in the Southwest (photograph by Jose Iniguez, USDA Forest Service). Figure 10 Mixed conifer forests are limited to a few isolated locations on north aspects in Saguaro National Park, but occur extensively across higher elevations in the GALWC. This scene is in the Mogollon Mountains taken prior to the 2012 Whitewater-Baldy fire that burned much of the landscape (photograph courtesy of USDA Forest Service). USDA Forest Service RMRS-GTR

12 Figure 11 Mixed conifer in the two study areas can be divided into (a) cool/moist and (b) warm/dry types. The cool/moist mixed conifer forests are distinguished by a general absence of ponderosa pine. The closed canopy and abundant coarse woody debris in the cool/moist mixed conifer forest make it ideal habitat for Mexican Spotted Owls (photographs by Calvin Farris, National Park Service). contain the same mix of associated species as the cool/ moist type but ponderosa pine is often still dominant or co-dominant (figure 11b). Cultural History The two study areas share many cultural history similarities. Prior to European settlement both areas were home to various Native American groups. The earliest known inhabitants in the greater GNF landscapes were the Mogollon culture. They were followed by the Pueblo Indian Cliff Dwellers who occupied the area in the 13 th and 14 th centuries and, like other native cultures, abruptly abandoned the areas for unknown reasons (Cordell 1997). In SNP the oldest known inhabitants to the general area were the Hohokam who, as farmers, mainly occupied the desert valley bottoms near reliable water sources. By the 13 th century, the Apaches began migrating to the Southwest from northern regions. Unlike prior inhabitants of the Southwest, the Apaches primarily hunted and gathered and thus utilized both valley bottoms as well as upland forested areas. Spanish settlers began moving into the Southwest in the 16 th and 17 th centuries (Clemensen 1987). Early Spanish explorers were mainly interested in finding gold and silver. By the 1600s, settlements developed around mining camps near what is now the GNF and around missions near the Rincon Mountains. These activities had a great impact on the land in the form of timber harvesting, which was needed both for construction and fuel. However, significant impacts were mostly confined to areas in close proximity to settlements. Moreover, through the 1800s the presence of 8 USDA Forest Service RMRS-GTR

13 Apache groups restricted Euro-American settlement in and around high elevation forests (Bigelow 1968). By 1856 both the Gila and the Rincon Mountains became part of the United States territory. Around the same time, the area saw the arrival of the railroad in 1880 and the capture of the last Apaches (Bigelow 1968). These two events allowed greater Euro-American settlement and brought about significant ecological changes throughout southwestern forests (Cooper 1961; Weaver 1951). By the close of the 19 th century, both areas had been declared part of the National Forest Reserve system and off limits to new settlement. By 1906 they became part of the newly created National Forest system. In 1924, due to the efforts of many including Aldo Leopold, the Gila Wilderness was set aside and became the first wilderness areas in the world (Roth 1990). In 1933 the Rincon Mountains were taken out of the National Forest system and designated a National Monument (Clemensen 1987). By 1976 the Saguaro Wilderness Area and the Gila Wilderness Area were officially designated with the passage of the Wilderness Act. The Aldo Leopold Wilderness Area was designated in 1980; in 1994 Saguaro was named the 52 nd National Park. HISTORY OF FIRE REGIMES AND MANAGEMENT With an abundance of ignition sources, receptive fuels, and frequent intra- and inter-annual drought, forested landscapes in the Southwest are some of the most fire-prone in the United States. These two study areas are no exception, yet fire occurrence has fluctuated a great deal over the past 120 years due to human influences. Thus, we divided the discussion of fire history in our study areas according to major human-caused changes in the fire regime. The Pre- Euro-American Settlement Era includes years prior to 1900, which is approximately when the arrival of Euro-Americans dramatically changed the role of fire in these landscapes. The Fire Exclusion Era encompasses the period of about 1900 to 1970 when fire was virtually eliminated from the landscape due to fire suppression and land use changes (heavy grazing, logging, etc.). The Fire Management Era begins around 1970 and continues into the beginning of the 21 st century. This represents the time when fires were reintroduced through prescribed fire and managed wildfire. We end the discussion with the present day as we enter the Megafire Era in the Southwest. This time period is characterized by the increasing incidence of large and intense wildfires concurrent with rising global temperatures and intensifying drought conditions (Stephens and others 2014; Westerling and others 2006). Pre-Euro-American Settlement Era (Prior to 1900) Prior to Euro-American settlement, the fire regime in each of these landscapes varied by vegetation type, reflecting the significant differences in rates of fuel accumulation and fire-season length. Within the pinyon-juniper woodlands, live and dead fuels accumulate relatively slowly due to limited moisture and high temperatures. Fire history studies in these woodlands in both SWA and the GALWC are lacking, but inferences can be made from studies in pinyon-juniper woodlands throughout the Southwest. Although the fire season is relatively long and ignitions are plentiful, widespread fires were likely infrequent historically. That is, the mean fire interval (MFI), or the average number of years between fires, was likely greater than 100 years due to slow rates of plant growth and the discontinuous nature of fuels (Romme and others 2009). When fires did occur, most were likely either very small (impacting one or a few trees) and with variable intensity, or large (several hundred acres) high-intensity crown fires. More frequent, low intensity fires likely occurred in some juniper savannahs or woodlands that supported a dense and continuous grass fuel bed on the order of every 20 to 30 years (Margolis 2014). Mid elevation pine forests have relatively high rates of fuels accumulation and a long fire season. Thus, they historically supported a fire regime dominated by frequent, low-severity surface fires prior to the 20 th century. In the MMSA in SWA for example, Farris and others (2013) found that the Natural Fire Rotation (NFR), or the time required on average to burn an area equivalent to size the entire study area, during the 19 th century was only 10 years. Between 1630 and 1900, a large fire (at least 25 percent of the study area) burned somewhere within MMSA every 7 to 8 years, on average (Baisan and Swetnam 1990; Farris and others 2013). The MFI within small stands or groups of stands ranged from 6 to 13 years in the GALWC prior to Euro- American settlement (Swetnam and Baisan 1996) (table 1). Fires of this frequency would have mainly consumed surface fuels and had relatively minimal impacts on the overstory. Most fires in both study areas occurred during the hot, dry arid foresummer between May and early July, prior to the onset of monsoonal moisture moving into the region (Abolt 1997; Baisan and Swetnam 1990). USDA Forest Service RMRS-GTR

14 Table 1: Summary of published pre-20 th century Mean Fire Return Intervals (MFI) in the Gila wilderness and Mica Mountain study areas (MMSA). Vegetation type sampled includes ponderosa pine (PIPO), mixed conifer (MC), and spruce-fir (SF). For blank values, the statistic either was not reported or not applicable because of stand size. Vegetation Elevation (feet) Fire-scar samples Area (acres) MFI all MFI 10% MFI 25% Time period Gila Wilderness PIPO/MC a 8,400-9, ~ PIPO a 8, ~ PIPO a 7,600-7, ~ PP b 7, MC/PIPO 8,640 1 n/a MC b 10, MC/PIPO b 8, MC b 8, SF/MC b 9, MC b 9, Mica Mountain PIPO c 8, d 7.3 e PIPO c 8, d 6.7 e PIPO c 8, d 6.9 e MC c 8, d 9.9 e PIPO/MC a 7,005-8, sites 6,870 g PIPO/MC f 7,005-8, sites 6,870 g a From Swetnam and Baisan b From Abolt c From Baisan and Swetnam d Approximate size of area individual trees were sampled from; estimated from maps in Baisan e Only includes intervals between fires that scarred more than one tree at a collection site. f From Farris et al g Size of the entire MMSA study area. Fuels also accumulate at relatively rapid rates in the mixed conifer forests, but the length of the fire season is shorter compared to ponderosa pine forests because snow can remain on the ground into the early summer months. Historical fire regimes in mixed conifer forests were highly variable ranging from relatively frequent low-severity surface fires in transition areas between pine and mixed conifer forests to infrequent high severity fires at higher elevations where mixed conifer forests transitions to spruce-fir. The extent of mixed conifer forests in the SWA is limited, covering less than 400 acres. These stands are generally found adjacent to pine forest but in cooler north-facing slopes and ravines and historically experienced surface fires of variable intensity at 15- to 30-year intervals (Baisan 1990; Iniguez and others 2009). In the GALWC, similar dry mixed conifer forests experienced a fire every 13 years on average (table 1), although individual intervals ranged between 3 and 30 years at the stand level (Abolt 1997; Baisan and Swetnam 1990). Although absent in the SWA, the GAWLC, also contains the wet/cool mixed conifer type on north-facing aspects that transition into spruce-fir forests. Winter snowpack tends to persist well into the summer in these forests and they are also the first to receive the summer rains, resulting in a relatively short fire season. As a result, these forests tended to experience relatively infrequent (every 30 to 400 years), mostly high severity fires (Abolt 1997; Margolis and others 2011). Although fires likely ignited somewhere on these landscapes on an annual basis, in most years fuels would have been too wet to support fire spread and fires would burn only a few acres. During very dry years, however, the wetter mixed conifer sites would support large standreplacing fires of 500 to 1,200 acres in size (Margolis and others 2011). 10 USDA Forest Service RMRS-GTR

15 Widespread fires across ponderosa pine and mixed conifer forests tended to be synchronized with interannual climate cycles that controlled broad-scale patterns of fuel production and drying (Swetnam and Baisan 1996). In the GALWC for example, Swetnam and Baisan (2003) identified at least 26 widespread fire years between 1605 and 1904 across a gradient of elevation and vegetation ranging from ponderosa pine to spruce-fir. Not surprisingly, most years with widespread fires were drier than average and most years with small fires were wetter than average. A closer examination of fire-climate relations reveals important lagging relationships during the years preceding fires: most of the largest fire years occurred during a drought year that was preceded by 1 to 3 years of significantly wetter/cooler conditions. These wet periods resulted in an abundance of fine fuels (especially grasses) across the landscape and set the stage for extensive fires when the next dry year occurred. Small fire years were preceded by relatively dry years that inhibited build-up of grasses and needles. These lagging relationships were weaker or absent in cool/wet mixed conifer forests dominated by woody fuels where drying was the most important factor (Swetnam and Baisan 1996, 2003). Fire Exclusion Era ( ) The cultural changes brought about by the arrival of Euro-Americans at the end of the 19 th century significantly altered the natural fire regime and ecology of these landscapes. After 1910, total annual area burned decreased dramatically in both the SWA and GALWC. For example, despite high rates of 20 th century burning compared to other pine-dominated forests in the West, the NFR in the MMSA increased from 10 years in the 19 th century to 31 years for the period 1937 to 2013 (Farris 2009; updated using data described in the methods section). A similar pattern occurred in GALWC fire history datasets (Swetnam and Baisan 1996). This was probably due to a number of factors but mainly the influx of intensive livestock grazing with the arrival of the railroad. Another contributing factor was that these areas were now under management of the U.S. Forest Service and active fire suppression was encouraged. The great numbers of livestock that moved into the Southwest with the establishment of the railroad significantly decreased grass cover and subsequently limited the capacity of the landscape to sustain frequent, low intensity fires (Cooper 1961; Weaver 1951). Furthermore, organized fire suppression began around 1910, and became increasingly effective with the use of mechanized equipment and aircraft in the 1940s and 1950s (Pyne 1997). During this Era, however, fire exclusion was certainly not ubiquitous across these landscapes; their relative remoteness and rugged terrain sometimes limited effective fire suppression. In particular, starting in the late 1940s and 1950s both areas saw the return Figure 12 Decadal fire perimeters for the Gila and Aldo Leopold Wilderness Complex and vicinity. Fire perimeters include managed and suppression fires for the entire period in addition to prescribed fires adjacent to the wilderness areas after Only fires that intersected wilderness lands are shown USDA Forest Service RMRS-GTR

16 Figure 13 Decadal fire perimeters in the Mica Mountain study area and vicinity. Fire perimeters include managed wildfire, suppression wildfire, and prescribed fires. Only fires that intersected wilderness lands are shown of some relatively large fires (figures 12 and 13). In the SWA, the first major fire in 1943 burned approximately 4,500 acres and by 1959 almost 8,000 acres had burned in total. What is particularly notable is that the major fires on Mica Mountain during this Era overlapped spatially, resulting in numerous reburns within a relatively short amount of time (figures 13). In the GALWC, several large fires occurred in the 1950s, including the McKnight Fire in 1951, which burned over 48,000 acres in the southern end of the Black Range. Although these were large fires compared to earlier decades in the Era, they were not as large as the frequent fires that burned across these landscapes prior to Still, between 1909 and 1969, a total of 134,000 acres burned in the GALWC; by 1970, approximately 25 percent of the wilderness area had experienced at least one fire since 1900 (figure 14). These mid-century fires reduced accumulated fuels and reduced tree densities, thereby making it more feasible to initiate managed wildfire programs in the 1970s (figure 15). Managed Fire Era ( ) The managed wildfire program in SWA began in 1971 following the development of a fire management plan approving the practice. It was the second national park in the country to adopt such a policy. Managed wildfire was first officially implemented in the Gila Wilderness in 1975 (Boucher and Moody 1998). The initial fire management plans, which were relatively conservative and simple by today s standards, described several Figure 14 Cumulative proportion of area burned at least once and at least twice over time in the Gila Aldo Leopold Wilderness Complex (GALWC) and Mica Mountain Study Area (MICA MTN). Data include all fire types. 12 USDA Forest Service RMRS-GTR

17 Figure 15 Evidence and legacy of mid-1900s fires. Despite aggressive fire suppression policies, several fires still grew large in the Saguaro Wilderness Area and Gila Wilderness in the mid-1900s. Panel (a) shows a major fire burning in the Gila Wilderness in Panel (b) shows the effects of a 1943 wildfire near Mica Mountain in the Saguaro Wilderness (photo taken before 1957) (photographs courtesy of: (a) the Forest History Society, Durham, N.C., and (b) the National Park Service (from Dodge 1957)). conditions required to consider managing a wildfire, such as specific calendar dates and a range of values for fire danger indices under which a wildfire could be managed. Additionally, early fire management plans in SNP highlighted the need for prescribed fire in the wilderness area to reduce fuels. With mechanical thinning not being an option in the Park or in wilderness areas, it was thought that prescribed fire in strategic locations would eventually allow for more use of managed wildfire. Prescribed fire was first implemented in the Park in Conversely, prescribed fire has not been implemented within the GALWC (although prescribed burning has occurred along the borders outside the wilderness), and managers have instead relied on natural ignitions to achieve their objectives. Because the programs were initially conservative with respect to allowable burning conditions, acreage burned by managed wildfire and prescribed fire was modest in the early years (~ ) (figure 16 and 17). From 1970 to 1989, a combination of all fires in the GALWC covered a total of 88,805 acres, 90 percent of those acres burned between 1985 and 1989 Figure 16 Area burned per year in the Gila and Aldo Leopold Wilderness Complex since The top panel shows annual area burned, and the bottom panel shows cumulative area burned in managed wildfires, suppression fires, and fires of unknown status (fires in which there was no information available in the database). Vertical dashed line indicates first year of managed wildfire in Prescribed fires have not occurred within these wilderness areas. The graph includes only fires greater than 100 acres. USDA Forest Service RMRS-GTR

18 Figure 17 Area burned per year in the Mica Mountain Study Area of Saguaro National Park since The top panel shows annual area burned, and the bottom panel shows cumulative area burned in managed wildfires, prescribed fires and suppression fires. Vertical dashed line indicates first year of managed wildfire in The graph includes only fires greater than 100 acres. *The 1994 Rincon Fire was initially a managed wildfire, but was converted to a suppression fire after it spread toward the Park boundaries. (figure 12). The largest managed fire during this period was the 1989 Shelley Fire that burned over 10,000 acres mostly within the wilderness boundary. Between 1971 and 1989, several managed wildfires occurred in the SWA, but the vast majority were relatively small (figure 17). The two exceptions were managed wildfires in 1972 and 1988 that burned 820 and 420 acres, respectively on Mica Mountain. These fires were particularly important because they reburned areas east of Reef Rock that had already experienced four overlapping fires in the 1940s and 1950s, and which would burn again multiple times in the future (figure 18). Additionally, the 1989 Chiva Fire was a suppression fire that started on Forest Service land but burned nearly 400 acres on Mica Mountain. In total more than 1,600 acres burned on in the SWA between 1971 and Large increases in prescribed fire and managed wildfire in both wilderness areas were not realized until the early 1990s and high rates of burning continued into the early 21 st century. There are numerous reasons that this pattern developed. By 1998, policy and fire management plans were updated to allow for managed wildfire outside of wilderness areas in the GNF and managers began utilizing managed wildfire in other parts of the GNF, mostly in the Black Range Ranger District. Increased Federal investment in fuels treatments and fire management in general, including the adoption of the National Fire Plan (2000) in the early 2000s, gave both units greater capacity to manage both Figure 18 View of the east side of Reef Rock on Mica Mountain in the Saguaro Wilderness showing a small lightning fire that was suppressed in The relatively open forest structure of ponderosa pine forests in the foreground reflects the influence of 5 to 9 repeated fires in this area since 1943 (photograph courtesy of the National Park Service). 14 USDA Forest Service RMRS-GTR

19 wildfire and prescribed fire. This included more funding for planning and implementing prescribed fire and access to Fire Use Modules (teams of fire professionals who specialized in implementing managed wildfire). Greater investment in prescribed fire ultimately provided more opportunities to utilized managed wildfire. For example, the Indian Peaks project in the GNF, in which 40,000 acres were treated with prescribed fire between 1999 and 2006, is seen as playing a pivotal role in allowing for the expansion of managed wildfire outside of the wilderness area in the Black Range. Prescribed fires on Mica Mountain in the SWA played similar roles. In addition, the accumulated experience with managed wildfire over decades allowed the organizations overall to be more comfortable and accepting of the practice. In the GALWC, more than 219,000 acres burned between 1990 and 1999 with over 60 percent of these acres burning as part of managed fires. By the end of the 1990s more than 40 percent of the area within the GALWC had burned at least once since 1910 and more than 20 percent had burned multiple times (figure 14). The new millennium brought even more fire and the increase in area burned can be summarized in three distinct types of fires. The first type was a continuation of managed wildfires inside the wilderness area, the largest of which was the 2003 Dry Lakes Fire Complex that burned nearly 114,000 acres. The second fire type was large managed wildfires outside the wilderness. Examples of such fires include the 2005 Black Range Complex (74,274 acres), 2005 Fork fire (14,869 acres), and the 2007 HL fire (10,784 acres). The third fire type was large-scale prescribed fires outside the wilderness. The largest of these fires was the 2003 Boiler fire (56,695 acres) that burned in pine forests just north of the Aldo Leopold wilderness. As in prior decades, there were also a small number of fires that burned within high elevation mixed-conifer forests. Patterns of severity of these fires varied, but fires in ponderosa pine forests tended to burn mostly with low severity while fires in mixed conifer forests tended to include more areas of high severity. For example, the 1992 Creel fire (17,240 acres), 1993 Brush fire (39,348 acres), 1993 Iron fire (5,825 acres) and 1997 Lilly fire (18,507 acres) all burned in ponderosa pine forest within the Gila wilderness and included less than 4 percent high severity within their perimeters (Iniguez 2014). These fires contained occasional patches of crown fire that created treeless patches as large as 300 acres. On the other hand the 1995 Bonner fire (26,380 acres) and the 1996 Lookout fire (14,200 acres) contained 18 percent and 11 percent high severity, respectively. These latter fires burned in higher elevation mixed-conifer forests; hence the mixed severity fire pattern was likely similar to historical fires in this landscape. In the SWA, a total of 5,374 acres burned on Mica Mountain between 1990 and 1999 (figure 17). More than half of this total was from the 1994 Rincon Fire (3,281 acres) that was initially a managed wildfire; it was later converted to a suppression fire when it exceeded resource capabilities and burned toward U.S. Forest Service lands where management options were more limited. While most acreage occurred in suppression status, we classified it as a managed fire for the purpose of this report because it would have been suppressed at <1 acre had policy not been in place to allow lightning fires to burn. The rest of the acreage burned during this period was from prescribed fires, most of which were implemented in areas that had been burned in the prior decades. By the end of the decade nearly 80 percent of the MMSA had burned at least two times since 1937 (figures 14). From 2000 to 2010 a total of 2,500 acres had burned on Mica Mountain, of which about half was in the form of managed wildfires and the other half as suppression fires. The largest of the suppression fires was the Helens II fire in 2003, which burned mainly on the north slope of Mica Mountain through mixed conifer stands that had not burned during the last century. Beginning in the mid- to late-2000s, a number of factors resulted in a decline in acreage burned in managed wildfire and prescribed fire in both the GAWLC and SWA (figure 16 and 17). This was driven in part by growing concern for the potential negative impacts of managed wildfire on valued resources, particularly under persistent drought conditions. Several suppression, managed, and prescribed fires during this period resulted in undesirable effects on habitat and populations of listed or sensitive species, including the Mexican spotted owl (Strix occidntalis lucida Nelson), the lowland leopard frog (Lithobates yavapaiensis Platz and Frost 1984), and the Gila trout (Oncorhynchus gilea (Miller)). In addition, some fires across the greater SWA spread into lower elevations in vegetation types not well adapted to fire; several saguaro cacti (Carnegiea gigantean (Engelm.) Britton & Rose), an iconic species and the namesake of the Park, were burned, along with other Sonoran desert species. There was also increasing concern about the spread of invasive grasses in the Sonoran desert, which might further encourage fire spread. In the GNF, there was growing concern for the impacts of fire on livestock range that was being impacted by severe drought. The capacity to implement prescribed fire in remote areas had also USDA Forest Service RMRS-GTR

20 Figure 19 Fire frequency in the Gila Aldo Leopold Wilderness Complex from 1909 to Most of the pine and mixed conifer forest in the GALWC have burned at least once; hence, most of the unburned areas are dominated by pinyon-juniper. Some areas in the interior of these two wilderness areas have burned as many as eight times since declined substantially toward the end of the 2000s; the cost of such projects had increased dramatically over the years and agencies directed most funds for fuels treatments to areas near the wildland urban interface. The last prescribed fire conducted in SWA was the 2010 Mica Mountain prescribed fire. Only 114 acres of the 423-acre project area were burned because conditions became too dry, yet the burn cost approximately $300,000. Finally, managers in both study areas had found difficulty in expanding managed wildfire and prescribed fire programs to the hard acres, the steeper terrain and mixed conifer forests with heavy fuel loading where fires have greater potential to burn severely and have detrimental impacts to critical resources. Such areas comprised most of the remaining long-unburned areas by the beginning of the 21 st century. Figure 20 Fire frequency in the Saguaro Wilderness Area in Saguaro National Park from 1937 to Most of the area within the Mica mountain study areas has burned at least one and some areas have burned as many as nine times since USDA Forest Service RMRS-GTR

21 Figure 21 View from Rincon Peak looking west toward the X-9 Ranch development. The steep terrain, accumulated fuels, and neighboring developments hamper fire managements in the Rincon Peak portion of Saguaro National park (photograph by Jose Iniguez, USDA Forest Service). Despite the successful reintroduction of prescribed fire and managed wildfire during this Era, there remain many areas within the SWA and the GALWC that have not burned in a century or more (figures 19 and 20). The most notable in the SWA includes much of the area on Rincon Peak, which historically had fewer natural ignitions and slightly longer fire return intervals than Mica Mountain, and has had no prescribed fire (Iniguez and others 2009). More conservative approaches to fire management (suppressing more natural ignitions) on Rincon Peak have been taken in part because limited access and heavy fuels provide few options for managers to prevent fires from spreading to neighboring private property (figure 21). In addition, there has been resistance from resource staff in managing fires in Rincon Peak where there are sensitive resources, such as spotted owl populations. By 2010, long-unburned areas in the GALWC included the upper elevation mixed conifer, aspen, and spruce-fir forests in the Mogollon Range, which by the end of the 20 th century, still hadn t burned in over 100 years. Managers often attempted to utilize managed wildfire in these areas, but when lightning fires occurred they often burned during the monsoon season and remained small before going out on their own. Many wildfires were also suppressed in this system because of concerns about detrimental fire effects on Gila trout and Mexican spotted owl habitat. In addition, unlike at the SNP, fire managers have not utilized prescribed fire within the GNF wilderness area. Thus, mixed conifer forests have experienced relatively low rates of contemporary burning and may be more susceptible to larger, high intensity fires compared to what these forests would have experienced historically. Megafire Era (2011 present) Very large and severe fires (~100,000+ acres) are projected to occur more frequently in the western United States as the climate becomes warmer and drier (Dillon and others 2012; Westerling and others 2006). There is overwhelming evidence that many types of fuels treatments can be effective in mitigating the severity of wildfires (Fulé and others 2012). However, the degree to which fuels treatments (including managed wildfire and prescribed fire) will be effective in making forests more resilient to wildfire under future climates remains an open question and an important topic in need of additional research (see Appendix B). After more than a quarter century of reintroducing wildland fire in the GALWC, this landscape has been tested by a series of large wildfires since These fires have highlighted the benefits and potential limitations of the managed and prescribed fire program under relatively warm and dry climatic conditions. In particular, the 2011 Miller Fire (88,835 acres), the 2012 Whitewater Baldy Complex (297,845 acres, currently the largest wildfire in New Mexico State history), and the 2013 Silver Fire (133,625 acres) burned during severe and prolonged drought conditions through forests previously treated with prescribed fire and managed wildfire and through long-unburned forests. USDA Forest Service RMRS-GTR

22 Figure 22 Map showing fire history and severity of the 2011 Miller Fire in the Gila National Forest. Most of the fire was low to moderate severity particularly within areas that had burned prior to Some of the areas that did experience high severity fires were pinyonjuniper stands that had not burned in the last century. Fire severity was derived from the Monitoring Trends in Burn Severity database ( mtbs.gov/). One thing that is evident from these fires is that the landscape mosaic created by previous prescribed, managed, and suppression fires provided options and opportunities to firefighters during suppression operations and that without that fire history, these suppression fires would have been much larger and more severe. For example, the spread of the Whitewater Baldy Complex (WWBC) was halted on the northeast by the 2006 Bear Fire and on the south by the Miller Fire. Firefighters were able to stop advancement to the north at the Eckleberger prescribed fire project area, an 18,000-acre burn on the Reserve Ranger District. Fewer resources were devoted to suppression on the eastern flank of the WWBC when it spread into Iron Creek Mesa, Jerky Mountain, and Lilly Mountain, where the WWBC was the third or fourth fire in the last 3 decades and was burning with low intensity. Similarly, the Silver Fire burned through previous fires with both low and high intensity, but it slowed and stopped to the north when it reached previous managed and suppression fires. The onset of monsoon conditions also helped stop the Silver Fire from spreading further. Managed wildfire and prescribed fire can increase the resiliency of these forests to wildfire, even under severe drought conditions. In maps that overlay fire history and burn severity (figures 22 and 23), one can see that most of the areas that had previously experienced managed wildfire reburned with low to moderate severity in the Miller Fire and the WWBC. Overall, more than 64,000 of the 88,000 acres burned in the Miller Fire experienced at least one fire in the previous few decades. Most of the area classified as high severity had not burned in over a century; this included pinyon-juniper woodlands, pine stringers, and riparian areas that burned as the Miller Fire approached the West Fork of the Gila River. The WWBC and Silver Fire resulted in severe effects mostly in upper elevation mixed conifer forests that were also long unburned. While the history of fires clearly impacted the spread and intensity of these suppression fires, it is also evident that previous fires were not always effective in mitigating fire severity. From the fire history and burn severity maps, it is clear that some areas classified as high severity had burned previously in managed wildfire (figures 22 and 23). Moreover, managers mentioned that all these fires seemed to result in fairly high tree mortality even as it passed through previously burned forests (figure 24). It is unclear, however, if the tree mortality is directly related to fire intensity, drought, or perhaps a combination of the two. Ultimately, the effects of subsequent fires likely depends on a large number of factors: how a fire burns into a stand, terrain and landscape context, fraction of total landscape treated, daily weather, seasonal drought, and fire management operations and tactics. More research into these and other factors that contributed to the high severity patches in these fires would further shed light on 18 USDA Forest Service RMRS-GTR

23 Figure 23 Map showing fire history and severity of the 2012 Whitewater Baldy Complex in the Gila National Forest. Fire severity was derived from the Monitoring Trends in Burn Severity database ( Figure 24 Aerial view of the 2011 Miller fire burning mostly as a low intensity wildfire within forested areas. The Miller fire also burned within and around nonforested patches (yellow/grass) created by prior more severe fires including the 2000 Bloodgood Fire and the 2003 Dry Lakes Complex. Patches of previous severe fires are now dominated by standing snags and grasses (photograph by U.S. Forest Service). the benefits and limitations of managed and prescribed fire under a changing climate (see Appendix B). The fire program in SWA has not been recently tested in the same manner, because total area burned has been very low over the past several years. We speculate that a similar pattern might emerge should a fire start under severe drought conditions. In the areas where previous prescribed fires and managed wildfires have occurred, subsequent fires should be more likely to spread with lower intensity compared to long-unburned areas. However, unlike the GAWLC where burned area is increasing, about 10 years has passed since the last major USDA Forest Service RMRS-GTR

24 fire in SWA. Given that prescribed burning has also declined substantially in the Park, it remains unclear if resource management objectives will continue to be maintained in this landscape in the future with managed wildfires alone. EFFECTS OF MANGED WILDFIRE AND PRESCRIBED FIRE Over the last several decades, many research and monitoring studies have been conducted in these study areas in an attempt to understand the effectiveness of managed wildfire and prescribed fire in meeting management objectives and the effects of the practices on critical resources. This section summarizes what has been reported in the literature and sheds light on areas in need of further research (see Appendix B). Much of the existing literature focuses on the effects of these practices on (a) fuels, forest structure, and potential fire behavior, and (b) wildlife populations and habitat. Fuels, Forest Structure, and Potential Fire Behavior The effects of reintroduced fire on fuels, forest structure, and potential fire behavior vary by vegetation type and historical fire regime. For example, widespread low intensity fires would likely have occurred every 10 to 30 years in juniper savannas (Margolis 2014). These areas have declined during the 20 th century as a result of grazing and fire suppression. However, pinyon-juniper woodlands historically made up a much greater portion of the landscape (Miller 1999). In the woodlands, fires historically either burned very small areas (1-2 trees) or burned large areas as crown fires. This also appears to be how managed wildfires currently behave in this system (figure 25). In two recent wildfires in the GNF (2006 Martinez and 2005 Johnson Fires), pinyon-juniper woodlands mostly burned as high intensity crown fires with almost complete overstory mortality (Hunter and others 2011). Even in pinyon-juniper woodlands that were classified as low severity in burn severity maps, often no evidence of fire was detected or if it was, it appeared that the fire impacted only 1 to 2 trees and logs in their immediate vicinity (Molly Hunter, personal observation of 2006 Martinez and 2005 Johnson Fires). Broadcast burning is not a common practice in these woodlands because fuels are generally not conducive to fire spread. Indeed, fire managers in the GNF often use pinyon-juniper woodlands as fuel breaks to slow or stop fire during prescribed fire and managed wildfire operations (Toby Richards, personal communication). However, recent wildfires have shown that many of these woodlands will support high intensity crown fire following a century of fire exclusion. While such fires may not be outside of the historical range of variability of fire behavior, it is unclear what the post-fire successional trajectories will be under current or future climates. Pinyon-juniper woodlands in other parts of the Southwest have shown the potential for novel successional trajectories due to spread of invasive species (Floyd and others 2006) and widespread mortality overstory induced by climate change (Breshears and others 2005). The severity of managed wildfire and prescribed fire in ponderosa pine forests has tended to be more moderate. Studies in both the GWALC and SWA have shown that managed and prescribed fires were effective at reducing surface fuels and density of small trees while Figure 25 High severity burn in pinyon-juniper woodlands in the 2011 Miller Fire. Many of these woodlands tend to burn during periods of high winds and drought resulting in higher severity fires (photograph by Jose Iniguez, USDA Forest Service). 20 USDA Forest Service RMRS-GTR

25 Figure 26 Many ponderosa pine stands that have burned one or more times have more heterogeneous, open stand structures with lower densities of small trees. Stands such as this one in the Gila Wilderness should now exhibit a reduced probability of high severity fires in near future (photograph by Jose Iniguez, USDA Forest Service). maintaining large trees (Holden and others 2007; Hunter and others 2011) (figure 26). This has resulted in stands that more closely resemble historical forest structure and are at lower risk of stand-replacing crown fire (Holden and others 2010; Hunter and others 2011). Areas that have previously burned in managed wildfires in the GAWLC tended to reburn with lower severity compared to previously unburned areas (Parks and others 2014). Furthermore, the cited studies show that those conditions can be effectively maintained after multiple managed wildfires have impacted an area (figure 27). One study showed that managed wildfire is slightly more effective Figure 27 Open pine forests in the (a) Gila-Aldo Leopold, and (b) Saguaro Wilderness following multiple surface fires. Open forests in Iron Creek Mesa in the Gila Wilderness have experienced three surface fires since 1970 (1985, 2003, and 2012) including burning most recently in the 2012 Whitewater-Baldy fire. In Mica Mountain, this ponderosa pine stand (b) did not burn between 1893 and 1942, but has since burned six times (two wildfires and four prescribed fires). Following a large increase in tree density during the fire exclusion gap in the early 1900s, the post-1942 fires significantly reduced tree density and increased structural diversity and understory herbaceous cover (photographs by Jose Iniguez, USDA Forest Service (a) and Calvin Farris, National Park Service (b)). USDA Forest Service RMRS-GTR

26 than prescribed fire at reducing fuels and potential fire behavior (Hunter and others 2011). However, this study focused on first entry prescribed fires, which managers tend to ignite with cooler prescriptions to minimize detrimental effects and improve control. While no study has specifically examined the effects of multi-entry prescribed fire, managers in both areas are confident that they have achieved fuels objectives with repeated prescribed fire (figure 28). Managed wildfires have been much less common in mixed conifer forests; when they have burned, they have tended to burn with higher severity compared to pine forests (Holden and others 2010). This may not be inconsistent with the historical fire regime, especially in the wetter, cooler mixed conifer types (Margolis and others 2011). In general, studies on the effects of prescribed fire and managed wildfire in mixed conifer forests in these areas are lacking in this region. However, Fulé and others (2004) found that a high intensity prescribed fire in mixed conifer forests in Grand Canyon National Park effectively reduced tree density and produced forest structures more closely resembling historical conditions. While studies on the effects of prescribed fire and managed wildfire on fuels, forest structure, and potential fire behavior have been studied in other parts of the country (for example, California), more studies in the Southwest are clearly needed (see Appendix B). Figure 28 Examples of managed and prescribed fires in the Gila National Forest and Saguaro National Park. The 2002 Cub Fire (a) is an example of a managed fire in the Gila National Forest. Panel (b) shows a ponderosa pine stand near the summit of Mica Mountain immediately following a prescribed burn in The burn was conducted as a maintenance burn in an area that had experienced four previous fires since 1943 (photographs by the U.S. Forest Service (a) and National Park Service (b)). 22 USDA Forest Service RMRS-GTR

27 Threatened, Endangered, and Sensitive Species In the last few decades, several local populations of the threatened Gila trout in the GALWC have been extirpated as a result of ash and debris flows following wildfires with large high severity patches on steep slopes (figure 29). Some examples include the 1989 Divide Fire, the 1995 Bonner Fire, the 1996 Lookout Fire and more recent wildfires (Brown and others 2001). Many populations of the Gila trout are geographically isolated, and opportunities for the fish to seek refuge and reestablish in recovered habitat after fires are limited. This has been exacerbated by low water levels during drought conditions. Thus, it has become common practice for fish to be removed from impacted streams when possible, and reintroduced when the habitat improves (Wood and Racher 2010). While widespread severe wildfire is known to be detrimental to the species, low intensity managed and prescribed fire may actually make Gila trout populations more resilient, especially if these fires result in reduced fuel loading and reduced risk of stand replacing wildfires in the future (Brown and others. 2001). Aquatic Habitat and Species Fires have also impacted aquatic species and habitat in the SWA. There are few perennial water sources in the SWA, and aquatic species rely in part on tinajas, or ephemeral pools that form in bedrock depressions. The lowland leopard frog, classified as a species of Figure 30 The lowland leopard frog is listed as a species of concern by the State of Arizona and is found in Saguaro National Park. The perennial pools that supply their habitat can fill with sediment following wildfire, but habitat and frogs have been found to return to some effected areas (photograph by Jim Rorabaugh, U.S. Fish and Wildlife Service). concern by the Arizona Game and Fish Department, is one of the species that relies on these pools (figure 30). Several tinajas filled with ash and sediment following the 1999 Box Canyon Fire and the 1989 Chiva Fire (Parker 2006). While water and frogs returned to some pools after a few years, some pools remained filled with sediment for up to 20 years. Severe fires have the greatest long-lasting impact on stream dynamics, sedimentation, and frog habitat. Thus there is concern that should fire spread through long-unburned portions of the SWA where fuel loading is uncharacteristically high, there could be further impacts on lowland leopard frog habitat. Mexican Spotted Owl Figure 29 The Gila trout is an endangered species found in streams within the Gila National Forest and is threatened by high intensity wildfire (photograph by R.C. Helbock, New Mexico Department of Game and Fish). The threatened Mexican spotted owl is adapted to fire-prone forests and uses a variety of forest conditions for feeding, nesting, etc. (figure 31). However, they rely particularly on closed forests with complex structure for nesting (USFWS 2012). These critical habitats exist both in pine forests within canyons and in denser mixed conifer forests. In the past, both managers and researchers have debated about using prescribed fire to reduce fuel loads within nesting areas as there was concern that fire could diminish key habitat features. However, some reduction in fuel loading is often desirable because stand replacing fires are seen as one of the biggest threats to the species (USFWS 2012). Extensive surveys have shown that low intensity prescribed fire USDA Forest Service RMRS-GTR

28 these sites particularly if patches of live forest remain (Willey 1998). However, there are fears because of low prey productivity and increased temperatures that the habitat will not be used by subsequent generation of Mexican spotted owl. Wildlife Habitat and Populations Figure 31 The threatened Mexican spotted owl is found in pine/oak and mixed conifer forests in both the Gila National Forest and Saguaro National Park. In some instances, managers have been successful in allowing spread of prescribed fire and managed wildfire in their habitat with minimal to beneficial effects. In other instances, managed wildfire and suppression fires have been detrimental to their habitat (photograph by Bill Radke, U.S. Fish and Wildlife Service). and managed wildfire have minimal short-term impact on Mexican spotted owl occupancy or reproduction (Jenness and others 2004). However, several recent wildfires, such as the 2002 Helen s II Fire in the SWA and the 2012 Whitewater-Baldy Complex in the GNF have burned severely in core owl habitat, particularly in mixed conifer forests (figure 32). Mexican spotted owls have high site fidelity so they will likely return to The fire programs in SNP and the GNF have had mostly positive or neutral impacts on non-listed wildlife species and their habitat. In the past, there have been concerns that frequent fire might deplete the snag population, which is a critical habitat feature for many wildlife species (Boucher and others 2000). Despite a decline in snag densities following a moderately intense fire in the Santa Catalina Mountains, they remained sufficient for wildlife habitat and bird populations remained stable (Ganey and others 1996). Snag densities also remained within recommended guidelines in areas that have burned multiple times in managed wildfire in the GALWC (Holden and others 2006). There is also evidence that some species in particular have benefited from managed wildfire programs. For example, the buff-breasted flycatcher (Empidonax fulvifrons) is common in Mexico but its numbers declined in the mountains of southern Arizona and New Mexico starting in the late 1800s; some research suggests that fire exclusion may be partly responsible for the trend. Surveys have found the buff-breasted flycatcher associated with areas that recently burned in surface fire in the Rincon Mountains while they were not found on other mountain ranges in the region that had little to no evidence of recent fire (Conway and Kirkpatrick 2007). In addition, abundance of many other forest birds found in southwestern mountain ranges has been found to be higher in recently burned areas (Kirkpatrick and others 2006). Figure 32 High severity burn patches following the 2002 Helens 2 Fire on the north slope of Mica Mountain. Historically, widespread fires burned across these north-slope warm/dry mixed conifer forests every years prior to 1886, but the area did not burn between 1886 and The north slope of Mica Mountain was habitat for Mexican spotted owls. Although the fire severity was relatively heterogeneous, the long term impacts on the owl habitat are still unknown (photograph by National Park Service). 24 USDA Forest Service RMRS-GTR

29 CURENT CHALLENGES TO IMPLEMENTING MANAGED WILDFIRE Their relative remoteness, inaccessibility, and history of frequent fire makes both of these study areas ideal for using fire as a restoration tool across broad landscapes. However, managers in both study areas still faced significant challenges managing wildfire and implementing large scale prescribed fire over the several decades of these programs. These challenges, which would likely be faced in other parts of the Southwest and the country, are described briefly here. In addition, we give examples of how managers in these study areas have worked to overcome these challenges. Threatened and Endangered Species There are several threatened and endangered species within SNP and the GNF that could be impacted by wildfire or prescribed fire: Mexican spotted owl, the lowland leopard frog, the Gila trout, and other species. Fire managers need to consider how their actions will affect these species. Managers often need to alter fire management practices during prescribed fire and managed wildfire to reduce potential threats to such species. Changed strategies and tactics could include altering the timing or intensity of fire, or restricting fire all together in sensitive areas. Managers in the GNF and SNP have overcome some of these challenges in particular by collaborating across disciplines and implementing rigorous monitoring programs. Managers in the GNF for example have been successful in allowing fire to spread through Mexican spotted owl nesting habitat in pine forests with minimal and even beneficial effects to the species. In other parts of the Southwest, the default management strategy has been to avoid these areas altogether, which may ultimately be detrimental to the species if their core habitat cannot be made resilient to wildfire. The success in the GNF is partly attributed to collaboration between the fire staff and the natural resource staff that work closely together and recognize the importance of meeting objectives for both fuels and wildlife. Before every fire season both fire and non-fire personnel participate in a meeting where they strategize about where they might allow for managed wildfire in the upcoming fire season. Because of this close working relationship, fire managers understand what kind of fire and effects are desirable in sensitive areas and they can help achieve those effects. As a result, fire managers can make more informed decision on where, when, and how to manage fires. SNP has invested a great deal in long-term monitoring to document the effects of fire on habitat and populations of lowland leopard frog to better understand the short- and long-term effects of fire on the species. Through these efforts, they have learned that leopard frog habitat can be severely degraded in the years following a high intensity wildfire and populations can subsequently disappear (Parker 2006). However, after several years frog populations have returned to some affected watersheds, indicating that refuge areas exist within the landscapes. Ultimately the monitoring results have provided further justification for the managed wildfire program; it has shown that fire management practices have likely increased resilience of frog populations in the long run. Public Support for Fire Programs Public support for fire programs is vital for their long-term success and there are many factors that can influence public opinion. For example, the public may not be supportive of fire programs if the inevitable smoke significantly impacts populated areas. However, the public is more likely to tolerate smoke if they recognize the benefits of less severe fires in the future. Other unintended consequences of prescribed fire and managed wildfire, such as escapes or impacts to recreation areas, can also influence public support. While some impacts to the public from fire are inevitable, minimizing impacts, engaging the public, and communicating the importance of fire can increase public support. Maintaining an active fire program is also important. Managers in both study areas have invested a great deal in public outreach, which ultimately has increased public acceptance of prescribed fire and managed wildfire programs. Managers in both study areas take a variety of outreach approaches ranging from interpretive signs, putting information officers at visitor centers or other public venues, maintaining detailed websites to inform the public about individual incidents and the overall fire programs, and conducting educational programs in schools. They also work directly with local media outlets, which is why one can often find wellinformed and thoughtful stories on the fire ecology and management of the areas in both local and national publications. It is critical to devote resources to public outreach year round and to frame different messages at different times (before, during and after fires). Maintaining an active fire program is also important; USDA Forest Service RMRS-GTR

30 having fire on the landscape each season (whether prescribed fire or managed wildfire) allows residents in the area to become accustomed to their effects, including smoke. A survey of local ranchers living near the GNF showed that many of these efforts have paid off; over two-thirds of respondents were supportive of the efforts of the GNF to manage wildfires and prescribed fires (Doug Boykin, personal communication). Risk of Escape There is always the potential for managed fires to spread beyond designated planning areas and impact sensitive resources or neighboring jurisdictions. This is a particular challenge in the SWA because it is a relatively small landscape surrounded by private, state, and Federal land managed by different agencies. Furthermore, the SWA boundaries are linear and do not follow any natural features that could be used to contain fires (figure 33). Any fire that starts in the SWA and grows to any measurable size is likely to spread beyond the Park s boundaries. Therefore, most large wildfires in SWA have eventually been controlled on some flanks or suppressed completely because they threatened the border with the neighboring Coronado National Forest (CNF) and private lands. This is less of an issue in the GALWC, which is much larger than SWA; however, managers in the GNF must still contend with fire spreading beyond the agency boundaries. Outside of the wilderness areas, there are a number of private land inholdings throughout the forest and managers are challenged with minimizing the impact to these areas. For example the 2006 Skates Fire eventually had to be suppressed to prevent it from threatening a nearby community. Managers in both areas have dealt with this challenge by collaborating with neighboring jurisdictions, both public and private. For example, SNP and the CNF developed a joint fire management plan, and a joint agency fire management officer oversees that plan. This has facilitated the sharing of resources and coordination of efforts across the agencies. In the GNF, managers worked with New Mexico State Forestry to develop agreements with private landowners in the area, mostly ranchers, to allow the GNF to manage wildfires on some private inholdings. These efforts allow for more effective and efficient management of wildfire as managers can take advantage of natural barriers for containment lines rather than trying to halt the spread of fire at fence lines. Invasive Species Spread of invasive species and their impact on fire spread patterns is becoming a significant challenge, particularly in SNP. Minimal fuels in Sonoran Desert habitats at lower elevations in the Park historically provide a barrier to fire spread from the higher elevation forests and woodlands. However, spread of invasive grasses such as buffelgrass (Pennisetum ciliare (L.) Link) at lower elevations has resulted in a more continuous fuel bed to support fire spread across these habitats Figure 33 Eastern boundary of Saguaro National Park. Saguaro National Park shares a boundary with the Coronado National Forest on the north, east and south, which often hampers the ability to manage fires due to the lack of natural barriers and differing levels of risk tolerance and management objectives. This image shows the eastern boundary, in close proximity to frequently burned ponderosa pine forests in the foreground) (photograph taken from Reef Rock looking southeast by Calvin Farris, National Park Service). 26 USDA Forest Service RMRS-GTR

31 Reintroducing Fire in Mixed to High Severity Fire Regimes Figure 34 Saguaro Cactus among a fuel bed of invasive grasses. Buffelgrass invasion of the lower elevations in Saguaro National Park has increased the connectivity of fuels across the landscape and can be detrimental to desert species that are not adapted to fire (photograph by Bethany Hontz, National Park Service). (figure 34). The potential for fire to spread in lower elevations could have significant impacts on many important native species in the Park that are not welladapted to fire, including the namesake saguaro cactus and many other species endemic to the Sonoran Desert. It may also result in more fires in higher elevations if continuous fuels allow fires to spread from lower elevation deserts (where human-caused ignitions are more common) to upper elevations woodlands and forests. Multiple efforts are underway at SNP to combat the spread of buffelgrass. This includes treating hundreds of acres each year with herbicide or manual pulling, engaging in research and monitoring to inform treatment regimes, and public outreach. Research has shown that management efforts have been effective in slowing spread of the grass (Hunter 2012). In addition, managers at SNP are heavily engaged in a model collaborative effort to raise awareness about the threat of buffelgrass and coordinate volunteer and control efforts across southern Arizona (see org, last accessed 6/10/13). Managers in both areas have had difficulty reintroducing fires to areas that historically burned with mixed to high severity, namely pinyon-juniper woodlands and mixed conifer forests. When managed wildfire has been attempted in these systems they often either burn out on their own after impacting limited acreage, or burn as large, high intensity crown fires. While this pattern may not be inconsistent with what one would have seen historically, the high intensity fires in particular can be undesirable because of their potential to impact critical resources (such as threatened and endangered species) and may be difficult to control. Thus, many wildfires in mixed conifer forests have been suppressed during dry conditions because of concern for impacts on species like Gila trout and Mexican spotted owl. While these species obviously persisted under the historical mixed to high severity fire regimes, they are more threatened by contemporary high intensity fires given their low population numbers. Plus, recent fires in mixed conifer forests (for example, WWBC) have resulted in high severity patch sizes that are larger than historical estimates, possibly due to greater fuel homogeneity at landscape scales (Ellis Margolis, personal communication). In particular, it appears the high severity patches are burning both in dry and wet mixed conifer whereas historically high severity were relatively small and limited in dry mixed conifer forests (figure 35). If fire is to be used as a tool in these systems, it is inevitable that some of it will be of high intensity and severity and it has been challenging to balance this fact with the protection of critical resources. The GNF has attempted to address this challenge by experimenting with ways to reintroduce fire to mixed conifer forests. For example, a project was planned to introduce prescribed fire to a watershed occupied by Gila trout to reduce fuels and increase the resiliency of the population in the face of future wildfires. Unfortunately the project area was burned in the WWBC before the prescribed fire could be implemented. Other recent projects include the 2012 Farm Flats Prescribed Fire, which covered approximately 2,000 acres of mixed conifer forests and was implemented in the fall to reduce fuels and protect a nearby community. Burning in the fall helped to achieve fuel reduction objectives while also keeping fire behavior within manageable levels. If fuels can be reduced in strategic locations to protect critical resources, then opportunities for use of some high USDA Forest Service RMRS-GTR

32 Figure 35 High severity from the 2012 Whitewater-Baldy fire. Although high severity fires are part of the historical fire regime in some mixed conifer forests, recent patches of high severity fire in the Gila National Forest may be larger due to fire exclusion, particularly in the dry/warm mixed conifer type (photograph by Jose Iniguez). intensity managed wildfire and prescribed fire may increase in other areas. LESSONS LEARNED Several factors facilitated the successful initiation of managed wildfire programs in the 1970s and their expansion over the last few decades. First the GAWLC and SWA represent relatively remote forested landscapes, making them ideal places to experiment with managed wildfire; while the SWA is not geographically remote, it is functionally remote due to limited access historically. Other important factors include the presence of vast fire-prone vegetation types, favorable climatic conditions, political environments that facilitated fire restoration, and the occurrence of mid-20 th century wildfires that reduced fuel loads in many areas. The environmental and sociopolitical settings of these landscapes differ significantly from many other land management units throughout the West. Thus, we don t necessarily expect the success of these programs to be replicated at similar scales everywhere using the same template. Nonetheless, there is much to be learned from these two programs that might be useful and applicable to other areas. Thus, we culminate this report by highlighting some of the key lessons that can be learned from these fire management programs. This list is by no means comprehensive or organized in order of importance. It simply represents recurring ideas and themes that emerged from the literature, the fire occurrence data, and discussions with managers throughout this project. Managed wildfire and prescribed fire can restore southwestern forests Managed wildfire, large-scale prescribed fire, and suppression fires create landscape fire mosaics that provide opportunities for fire managers Collaboration across disciplines and agencies is imperative Preparation and public outreach is critical Managed Wildfire and Prescribed Fire Can Restore Southwestern Forests Fire has been an effective restoration tool in the SWA and GAWLC to reduce fuels, restore forest and landscape structure, and reduce the potential for intense wildfire (particularly in the ponderosa pine dominated forests). Overall, both wilderness areas have burned considerably more during the 20 th century than most other forested landscapes in the western United States. Within the SWA and GALWC there are extensive areas that have burned three or more times since the inception of the managed wildfire programs, 28 USDA Forest Service RMRS-GTR